Rob De Langhe
rob.delanghe at telindus.be
Wed Jun 18 04:55:32 EDT 2003
got some replies, all in a straight line to the answer: the "snmpd" of
net-snmp package does active monitoring of aspects of the local host, but
does not generate SNMP-traps itself when a problem is detected.
Instead it sets MIB variables to indicate an error condition with one of the
parameters that it monitors.
For example, specifying in the /usr/local/share/snmp/snmpd.conf file:
disk / 80%
then (sure, after having started the "snmpd")
snmpwalk -c <ro-communitystring> localhost dskTable
reports error conditions like
UCD-SNMP-MIB::dskErrorFlag.<n> = INTEGER: <X>
if <X> is 0, there is no problem with that monitored disk
if <X> is 1, the threshold for that disk is exceeded; the variable
UCD-SNMP-MIB::dskErrorMsg.<n> will contains the message saying (in this case
obvious) that capacity threshold has been reached/exceeded.
Thanks to John Surveyor and Samier Kesou.
From: Rob De Langhe
Sent: Tuesday, June 17, 2003 7:37 AM
To: 'sunmanagers at sunmanagers.org'
anyone out there who successfully got the net-snmp agent "snmpd" to verify
the local system's health according to its config file contents?
I told it to monitor as example one filesystem's capacity, one file's size,
one missing process, and gave it the name of an SNMP-trap server host. But
not a single SNMP-TRAP packet (port 162/udp) was sent from the machine.
Any clue, hints or tips?
sunmanagers mailing list
sunmanagers at sunmanagers.org
More information about the summaries